The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from external NFC tags, which represents an indirect prompt injection attack surface.
Ingestion points: Data enters the agent context via the readerSession(_:didDetectNDEFs:) and tagReaderSession(_:didDetect:) delegate methods in SKILL.md, and through NSUserActivity restoration handling in references/background-and-entitlements.md.
Boundary markers: The provided code examples do not utilize boundary markers or explicit instructions to treat the ingested tag content as data rather than instructions.
Capability inventory: The skill enables the agent to write to NFC tags via writeNDEF and sendCommand, and to process URI payloads that can trigger application routing or web navigation.
Sanitization: No sanitization or validation logic is demonstrated for the NFC payloads before they are used in the application flow.

core-nfc